Question Formation in Québec* Martin Elsig and Shana Poplack 1

Transplanted dialects and language change: question formation in Québec*

Martin Elsig and Shana Poplack

1 Introduction

Variability in question formation is a well-documented feature of French syntax. In yes/no questions, five distinct variant forms have been competing for centuries: inversion of (clitic) subject and verb (INV), as in (1), complex inversion (C-INV) (2), rising intonation (INT) (3), phrase-initial interrogative particle est-ce-que (ECQ) (4), and its post-verbal counterpart tu (TU) (5).

(1) As-tu (INV) déjà parlé avec un vrai Français de France là? (XX.105.2768)1 ‘Have you ever spoken to a real Frenchman from France?’ (2) Et le roi est-il (C-INV) icitte? (XIX.036.3932) ‘And the king, is he here?’ (3) Ah, toi tu restes pas (INT) avec tes parents? (XX.112.1819) ‘Oh, you don’t live with your parents?’ (4) Mes bombes est-ce que (ECQ) je les largue ici? (XX.078.1502) ‘My bombs, do I throw them here?’ (5) Tu vas-tu (TU) être plus marié oubedonc moins marié? (XX.079.1471) ‘Are you gonna be more married or less married?’

Empirical studies of European varieties report that the variability illustrated in (1-5) has resolved itself in favor of INT (3), with ECQ persisting as a minor contender. INV, once the quintessential interrogative marker, is now

* The research reported here is part of a larger project entitled Confronting pre- scription and praxis in the evolution of grammar. We gratefully acknowledge the support of the Sonderforschungsbereich Mehrsprachigkeit to Elsig, and that of the SSHRC and the Killam Foundation to Poplack. Poplack holds the Canada Research Chair in Linguistics. 1Codes refer to corpus (XIX =Récits du français québécois d’autrefois [Poplack & St-Amand 2002]; XX =Corpus du français parlé à Ottawa-Hull [Poplack, 1989]; XVII =Corpus of 17th-century popular French plays), speaker and line number. Ex- amples are reproduced verbatim from audio recordings or plays.

U. Penn Working Papers in Linguistics, Volume 12.2, 2006 2 MARTIN ELSIG & SHANA POPLACK restricted to literary use. The spread of TU is said to have been blocked by stigma (Behnstedt 1973:32, Foulet 1921:271-272), while C-INV has disappeared altogether. In Québec French, on the other hand, not only are the “ex- tinct” variants thriving, but as we shall see, they each fulfill a well-defined function. As a result, the Canadian system of question formation appears structurally more complex than that of its source, the opposite of what is expected of transplanted dialects (e.g. Britain 2004). In this paper, we investigate the continuing evolution of question formation by tracing the variable expression of yes/no questions before and after the French settlement of Québec during the 17th century.

2 Data and Method

The corpora on which our analyses are based, displayed in Table 1, are par- ticularly well-suited to this endeavor.

Century Source Time span

20 Corpus du français parlé à Ottawa-Hull (spkrs b. 1898-1965) Poplack (1989)

19 Récits du français québécois d'autrefois (spkrs b. 1846-1895) Poplack & St-Amand (2002)

17 17th-century popular French plays (1629-1663)

16-20 Répertoire historique des grammaires du français (1530-1998) Poplack, Jarmasz, Dion & Rosen (ms)

Table 1: Data sources

Two represent vernaculars spoken in Québec in the 19th and 20th centuries. A corpus of 17th-century popular plays and a compilation of normative gram- mars dating from 1530 to the present are complementary diachronic sources representing a benchmark before the language was transplanted. We make use of the prescriptive tradition to date the variant forms, assess their institutional acceptance, and most important, ascertain the factors conditioning their selection (Rosen 2002, Poplack, Jarmasz, Dion, & Rosen (ms); Poplack & Dion 2004; Poplack, Dion, Jarmasz & Leblanc 2002). Real-time analysis spanning several centuries will help pinpoint the locus and time of change, if any. TRANSPLANTED DIALECTS AND LANGUAGE CHANGE 3

3 Variable Context

From each of the usage corpora we extracted every non-rhetorical question requiring a yes/no answer, noting the variant selected for each, excluding all others (e.g. wh-questions, as in (6), fixed expressions (7), non-sentential questions (8), imperatives (9), rhetorical questions (10), echo questions (11), and interrogative tags (12)).

(6) Il dit « Pourquoi tu as pas tué l’ours blanc? » (XIX.036.2670) ‘He says « Why didn’t you kill the white bear? »’ (7) Hey, ça fait longtemps, tu sais? (XX.096.79) ‘Hey, it’s been a long time, you know?’ (8) Les pattes rondes ça? (XIX.043.2452) ‘The round paws, there?’ (9) Allons aux moutons? (XIX.054.1893) ‘Shall we get to the point?’ (10) Tu as peut-être déjà vu ça ces chartières là ou entendu parler… ? (XIX.18.820) ‘You may have already seen those window bars, or heard of them… ?’ (11) “Ah moi j'aime pas ça”. “Tu aimes pas ça? (XX.103.398) ‘ “Ah, I don’t like that”. “You don’t like that?”’ (12) Ils grasseyent eux-autres à Montréal, hein? (XX.089.1725) ‘They have a guttural R in Montreal, eh?’

Table 2 compares variant distribution in contemporary Québec and European French.

Quebec French European French T his Fox Pohl Terry Ashby Söll Coveney study 1982 1965 1970 1977 1982 2002 % % % % % % % Intonation 35 36 86 86 80 91 79 -tu 33 34 0 - - - - Inversion 26 29 0 11 9 1 - Est-ce que 6 1 14 3 11 8 21 Complex-inv. ------Total 776 871 816 3016 130 452 180

Table 2: Distribution of variants in Québec French and European French 4 MARTIN ELSIG & SHANA POPLACK

INT is basically the only viable variant in Europe; in Canada, there are three. Why should speakers select one rather than another? Most scholars who have addressed this issue believe the variants differ pragmatically, conveying nuances like emphasis, doubt, astonishment, weak curiosity, etc. (e.g. Behnstedt 1973, Coveney 2002, Pohl 1965). We are less sanguine about our ability to identify such readings in discourse, let alone correlate them with specific variants. Instead, we examine the role of linguistic factors (e.g. subject type, and frequency, form, semantics, syllable length and lexical identity of the verb, as well as polarity). Most of these, along with speech style, have been invoked for centuries as explanatory of variant choice (e.g. Ashby 1977, Behnstedt 1973, Coveney 2002, Dewaele 1999, Pohl 1965, Söll 1982, Terry 1970). We analyze their combined effect using Goldvarb 2001 (Rand & Sankoff 1990), which enables us to contextualize the role of the variants within the system, with a view to elucidating its evolution over time.

4 Results

Turning first to contemporary Canadian French, a first important finding (Fig. 1) is that negative polarity is overwhelmingly expressed by INT, as in (3). Ensuing analyses thus deal only with affirmative questions.

Figure 1: Distribution of variants according to polarity

TRANSPLANTED DIALECTS AND LANGUAGE CHANGE 5

Table 3 displays four independent variable rule analyses of the other factors selected as significant to variant choice.

Variant Inversion -tu Est-ce que Intonation

Total N 205 247 50 186 Corrected mean: .30 .36 .07 .27 Subject type 2nd person 100% .36 .37 [ ] Others 0% .70 .69 [ ] Verb form Synthetic .47 [ ] [ ] [ ] Periphrastic .64 [ ] [ ] [ ] Verb semantics "Cognitive" .77 .43 - [ ] Others .36 .54 100% [ ] Verb syllables Polysyllabic .64 .33 .63 [ ] Monosyllabic .45 .57 .45 [ ] Verb frequency Frequent (21+) [ ] [ ] [ ] [ ] Mid (6-20) [ ] [ ] [ ] [ ] Rare (1-5) [ ] [ ] [ ] [ ] Style Careful .44 .46 .65 .52 Casual .61 .58 .24 .45

Table 3: Variable rule analyses of factors selected as significant to variant choice: 20th century (affirmative tokens only; brackets indicate non- significant factors)

The table shows that the major task of INV is to form direct questions, as in (13), while questions involving other subjects tend to be formed with the interrogative particle TU, as in (14).

(13) Bien j’ai dit, es-tu (INV) fou toi? (XX.112.1980) ‘So I said: “Are you nuts?”’ (14) Bien j’avais-tu (TU) de l’air niaiseuse? (XX.117.2122) ‘Well, did I look silly?’ 6 MARTIN ELSIG & SHANA POPLACK

Other contexts favoring INV, albeit to a lesser degree, include verbs of cognition, as in (15), and polysyllabic verbs (16). Here INV is complementary to TU, which in turn behaves like a default variant. We also note that verb frequency, claimed by many (e.g. Ashby 1977, Behnstedt 1973, Dewaele 1999, Pohl 1965) to explain why INV persists at all, has no effect. (Nor does lexical identity, though not shown here). We conclude that INV remains productive in contemporary Canadian question formation.

(15) Maintenant, comprenez-vous (INV) que pour le minéral je peux le localiser maintenant? (XX.082.2979) ‘Now, do you understand that for the mineral, I can locate it now?’ (16) Mangeriez-vous (INV) votre père? (XIX.004.1654) ‘Would you eat your father?’

The variants also have strong stylistic connotations, and these too are the opposite of those reported for Europe: In Canada, the rare ECQ (17), and to a lesser extent, INT (18), denote formality, while the other variants are rele- gated to casual speech (19-20).

(17) Excusez, est-ce que (ECQ) je peux le regarder? (XX.091.270) ‘Excuse me, may I see it?’ (18) Vous êtes correcte comme ça ? (INT) (XX.119.2053) ‘Is everything all right this way?’ (19) Penses-tu (INV) que j’étais faite comme un boeuf? (XX.009.1478) ‘Do you think I was built like an ox?’ (20) Ouais, ça a-tu (TU) du sacre bon sens ? (XX.84.1867) ‘Yeah, does that make any goddamn sense?’

We may summarize the main functions of the interrogative markers in 20th century Québec French as follows: a) negative polarity questions are expressed with INT, b) INV is specialized for direct questions, c) ECQ is a hyperstyle marker, and d) TU assumes most of the remaining (non- specialized) work of question formation. This pattern is substantively differ- ent from what is reported for European French (Table 2), where INT is the default variant, with only a little support from ECQ, while TU, a frontrunner in Québec French, is practically nonexistent. This raises the question of how the Canadian system arose. Since European French is the source language, it would be reasonable to assume that Canadian French is the innovator.

4.1 Development of the French interrogative system

But a brief review of the historical record reveals that all of the variants have TRANSPLANTED DIALECTS AND LANGUAGE CHANGE 7 a long and venerable history, linked to the Old French shift from VS to SV word order and ensuing efforts to rout out remaining inversions. This paved the way for the incursion of other forms, all of which had the virtue of re- establishing, the desired SV order. As far back as Middle French, INV co-existed with C-INV (Foulet 1921, Roberts 1993), which provided a solution to the problem of question formation with nominal subjects, since it contains both SV and VS word order. TU is widely believed to have originated through reanalysis of C-INV. But for reasons that are still not entirely clear, it was never accepted in polite discourse. Eventually, on analogy with the qu-est-ce que ‘what is it that’ paradigm for wh-questions, the construction ECQ emerged, grammaticizing to interrogative particle by the 16th century. In contrast to TU, this variant came to be explicitly ratified by the Académie Française, especially in of- fending contexts involving 1st conjugation and monosyllabic verbs (Vaugelas 1880/1884). INT, though perhaps the oldest (and currently the majority) variant in France, was never viewed as a serious contender. Summarizing, according to historical accounts, the prototypical interrogative variant, INV, was ousted due to loss of the VS word order it instan- tiates. It is said to have disappeared first from the most salient or problem- atic contexts (e.g. subject NPs, first person pronouns, 1st conjugation, and other monosyllabic verbs), and persisted where it was most entrenched: frequent verbs and collocations (Behnstedt 1973, Coveney 2002, Dewaele 1999, Pohl 1965). Ensuing analyses test these hypotheses. INV’s competitors, all of which reinstated the desirable SV order, infiltrated the system to vary- ing degrees, due, at least in part, to differences in institutional and social acceptance.

4.2 Question formation in 17th century French

Having reviewed the reported state of the language at the time of the French colonization of Québec in the 17th century, we now investigate the extent to which the scenario outlined above captured contemporaneous usage, as instantiated in the works of Corneille, Molière and Richer, popular play- wrights of the time. We make no claims about the extent to which rates of variant use reflect the speech of either actors or audience. But the structure of their variable selection, as revealed by the constraint hierarchies associated with each, can yield valuable information on their trajectory of development. This will be our focus in this section.

8 MARTIN ELSIG & SHANA POPLACK

Variant Inversion Intonation Pronominal Complex Total N 502 57 205 Corrected mean: .65 .07 .27 Subject type 2nd person [ ] - .59 Others [ ] 100% .42 Verb frequency Frequent (21+) .56 [ ] .46 Mid (6-20) .42 [ ] .52 Rare (1-5) .38 [ ] .60 Verb form Synthetic .51 [ ] .48 Periphrastic .40 [ ] .64 Verb semantics "Cognitive" .58 .34 .43 Others .45 .61 .55 Verb syllables Polysyllabic [ ] .39 [ ] Monosyllabic [ ] .58 [ ] Style Careful [ ] [ ] [ ] Casual [ ] [ ] [ ] Table 4: Variable rule analyses of factors selected as significant to variant choice: 17th century (affirmative tokens only)

Table 4 displays two independent variable rule analyses of the contribu- tion of factors to variant choice in 17th century plays. The corrected means indicate that INV, mostly pronominal, was still by far the majority variant of the time; the rest is basically made up of INT. We first note that subject type, the most important predictor of INV in 20th century French, was not significant in the 17th. This is because at the time, INV still occurred with all pronominal subjects, including the undesir- able je, a full 82% of which were inverted. This is illustrated in (20).

(20) Vous offensé-je (INV) en parlant de la sorte? (XVII.M.154.781) ‘Do I offend you by speaking this way?’

2nd person questions favored INT, and questions involving subject NPs TRANSPLANTED DIALECTS AND LANGUAGE CHANGE 9 were formed with C-INV or, to a lesser extent, INT. At this stage INV in yes/no-questions was already vanishingly rare with NPs, but still occurred freely with all pronominal subjects. This is consistent with the observed frequency effect: INV was already favored in frequent verbs, foreshadowing its eventual recession. The 20th-century tendency for INV to occur with “cognitive” verbs was already in place.

Although the Académie prescribed ECQ to replace INV in first person subjects, especially with monosyllabic verbs, the form was considered too colloquial throughout the 17th century to be admitted to writing (Foulet 1921). This may explain why we found so few of them (N = 14) in the plays. In this context, it is perhaps not surprising that number of syllables was not significant. But we can already detect a clear preference for questions involving monosyllabic verbs to occur with another variant: C-INV.

Summarizing, in 17th century (approximations of) speech, INV was still used productively with a variety of subject pronouns and verbs, though there were already harbingers of the current situation, since questions involving NPs, monosyllabic and infrequent verbs tended not to be inverted, all testify- ing to the gradual restriction of this variant.

4.3 The trajectory of variant choice over three centuries

Table 5 summarizes the trajectory of variant choice over the three centuries we studied, as illustrated by the relative contributions to their selection over the duration. Despite rate changes (observed in the corrected means), in all cases but one, the hierarchy of linguistic constraints is the same from the 19th to the 20th centuries. This suggests that any changes to the system must have pre- dated this time. Even the clear stylistic associations noted earlier for contemporary French were firmly in place by the 19th century: INV -- literary for Europeans -- was already favored in casual contexts, as was, more predicta- bly, TU. INT, and especially ECQ (both colloquial in European French) had become formal variants. But from a linguistic perspective, the system has remained essentially unchanged. For example, although C-INV disappeared, its associated conditioning was transferred to TU. Constraints on TU are now largely mirror-images of those on INV, consistent with its emerging role as majority variant: Where INV is favored, TU is disfavored. Even the differences in the contributions of subject type and verb frequency seem to be apparent only. Recall that in 17th century France INV could still occur freely across the entire pronominal paradigm, explaining why subject type 10 MARTIN ELSIG & SHANA POPLACK was not selected as significant. Two centuries later, INV was highly pro- moted by 2nd person pronouns, a constraint which has now become categori- cal. But despite its productivity with pronominal subjects, INV was receding from rarer contexts. So even though frequency appears from Table 5 to have had no effect on variant selection since the 19th century, we suggest that this is what in fact underlies the retention of INV.

Variant Inversion -tu Est-ce que Intonation Pronominal Complex Century 17 19 20 17 19 20 17 19 20 17 19 20 17 19 20 Total N 502 289 205 57 2 0 0 157 247 10 3 50 205 171 186 Corrected mean: .65 .47 .30 .07 - - - .25 .36 .01 - .07 .27 .28 .27 Subject type tu/vous [ ] .77 100% - - - - .31 .36 - - .37 .59 [ ] [ ] Others [ ] .06 0% 100% - - - .88 .70 - - .69 .42 [ ] [ ] Verb frequency Frequent (21+) .56 [ ] [ ] [ ] - - - [ ] [ ] - - [ ] .46 [ ] [ ] Mid (6-20) .42 [ ] [ ] [ ] - - - [ ] [ ] - - [ ] .52 [ ] [ ] Rare (1-5) .38 [ ] [ ] [ ] - - - [ ] [ ] - - [ ] .60 [ ] [ ] Verb form Synthetic .51 .53 .47 [ ] - - - [ ] [ ] - - [ ] .48 [ ] [ ] Periphrastic .40 .37 .64 [ ] - - - [ ] [ ] - - [ ] .64 [ ] [ ] Verb semantics "Cognitive" .58 .64 .77 .34 - - - [ ] .43 - - - .43 .38 [ ] Others .45 .44 .36 .61 - - - [ ] .54 - - 100% .55 .56 [ ] Verb syllables Polysyllabic [ ] [ ] .64 .39 - - - [ ] .33 - - .63 [ ] .42 [ ] Monosyllabic [ ] [ ] .45 .58 - - - [ ] .57 - - .45 [ ] .54 [ ] Style Careful [ ] .27 .44 [ ] - - - .19 .46 - - .65 [ ] .80 .52 Casual [ ] .52 .61 [ ] - - - .54 .58 - - .24 [ ] .47 .45

FactorsTable not5: V selectedariable as ru significant:le analyses of factors selected as significant to variant Subject typechoice: x17th, 19th and 20th -cen-tur-ies (affirmat-ive- tokens only) x x Verb frequency x x x - - - x x - - x x x VerbB formut why should INV be rexstric-ted- to- 2ndx persxon s-ubj-ectsx? Figure 2x shoxws tVerbhat t semanticshese have always accounted -for- m-ore xquestion-s t-han any other graxm- mVerbatic syllablesal person. xWe cxan assume th-at t-he -Frexnch brou-ght- to Canaxda also coxn- tStyleained a dispropxortionate numxber- of- su-ch contexts-, ex-plainingx the current persistence of INV here.

TRANSPLANTED DIALECTS AND LANGUAGE CHANGE 11

Figure 2: Distribution of questions according to grammatical person of addressee

The only real departure from 17th-century French involves negative polarity: it was strongly associated with INT by the 19th century in Québec. There is no evidence of this effect in the earlier materials.

5 Discussion

We may now return to the initially puzzling distributions displayed in Table 1. The Québec interrogative system, for all its differences from its contemporary European counterparts, appears to be a direct descendant of the one in use in 17th century France at the time of colonization. That system still prominently featured INV. Its total demise was blocked, not by entrench- ment in frequent verbs or collocations, but rather, by the preponderance of its favored contexts of occurrence: direct questions. C-INV, already moribund in the 17th century, disappeared altogether, but not before it transferred its functions to its alter ego TU. We know that TU (in its variant form [ti]) was (and is) widespread in many regions of France, though heavy institutional stigma curtailed its use. That stigma either failed to accompany the form to Canada, or disappeared thereafter. Table 6 confirms that not only is TU clearly on the increase among the young, but it is also favored by women. ECQ, still too rare in the 17th century, failed to gain a true foothold in Québec. This paved the way for its eventual conversion into a hyper-formal, upper-class variant, never used by speakers under 35. INT, interrogative marker par excellence in European French, somehow associated itself with negative questions. This is the only devel- 12 MARTIN ELSIG & SHANA POPLACK opment for which we have found no precursor in the history of the language, though it is also attested in contemporary European varieties (Coveney 2002:212-213 citing Borillo 1979, Mosegaard Hansen 2001, Söll 1971, Terry 1970).

Variant Inversion -tu Intonation Est-ce que Total N 205 247 186 50 Corrected mean: .30 .36 .27 .07 Sex Female [ ] .55 [ ] [ ] Male [ ] .45 [ ] [ ] Age 35+ .56 .42 [ ] 100% 15-34 .37 .69 [ ] - Socioeconomic class Upper middle [ ] [ ] [ ] .85 Working [ ] [ ] [ ] .16

FactorsTable 6 :notVar selectediable rul eas an significant:alysis of extralinguistic factors selected as significant to variant choice: 20th-century (affirmative questions only) Sex x x x

AgeThe distribution of these same interrogative variantsx in Europe is con- sSocioeconomicidered so com pclassartmentalized xthat their vaxriable selecxtion is said to be achieved through code-switching. INV -- simple and complex -- is a feature of Standard French grammar, while INT and ECQ are restricted to “français populaire” (Kaiser 1996, De Wind 1995). For Canadian French, no such analysis is required, since all but one of the variants continue to be impli- cated in question formation. Moreover, our analysis suggests that the contemporary Canadian interrogative system is a rather faithful representation of the system brought over from France. Aside from some rate differences involving the two major contenders, little has changed since the 17th century. The spectacular changes have taken place in France, for reasons no one has yet elucidated. Given the remarkable continuity of the linguistic conditioning of question formation, we may surmise that, as in Canada, it is the social embedding which has driven the change.

References

Ashby, William (1977). Interrogative forms in Parisian French. Semasia, 4, 35-52. TRANSPLANTED DIALECTS AND LANGUAGE CHANGE 13

Behnstedt, Peter (1973). Viens-tu? Est-ce que tu viens? Tu viens? Formen und Struk- turen des direkten Fragesatzes im Französischen, Tübingen: Narr. Borillo, A. (1979). La négation et l'orientation de la demande de confirmation, in: Langue Française 44, 27-41. Britain, David. (2004). Geolinguistic Diffusion of Language. In Ammon, U., N. Dittmar, K.J. Mattheier and P. Trudgill (eds), Sociolinguistics. An International Handbook of the Science of Language and Society. Berlin: Walter de Gruyter. 2nd edition. pp. 34-48 Coveney, Aidan (2002). Variability in Spoken French. A Sociolinguistic Study of Interrogation and Negation. Bristol & Portland: Elm Bank De Wind, Maarten (1995). Inversion in French, Ph.D. dissertation, University of Groningen, Groningen. Dewaele, Jean-Marc (1999). Word order variation in French interrogative structures, in: ITL Review of Applied Linguistics 125-126, 161-180. Foulet, Lucien (1921). Comment ont évolué les formes de l'interrogation, in: Roma- nia XLVII, 243-348. Fox, Cynthia (1989). Syntactic Variation and Interrogative Structures in Québécois. Ph.D. dissertation, Indiana University, Bloomington Kaiser, Georg A. (1996). V2 or not V2? Subject-verb inversion in Old and Modern French interrogatives, in: Brandner, Ellen; Ferraresi, Gisella (ed.), Language Change and Generative Grammar, Opladen: Westdeutscher Verlag, 168-190. Mosegaard Hansen, Maj-Britt (2001). Syntax in interaction. Form and function of yes/no interrogatives in spoken standard French, in: Studies in Language 25 (3), 463-520. Pohl, Jacques (1965). Observations sur les formes d’interrogation dans la langue parlée et dans la langue écrite non-littéraire. In Straka, Georges (ed.) Linguisti- que et philologie romanes. Actes du Xe Congrès International de Linguistique et Philologie Romanes, vol. 1, Paris: Klincksieck : 501-513. Poplack, Shana (1989). The care and handling of a mega-corpus, in: Fasold, Ralph; Schiffrin, Deborah (ed.). Language Change and Variation. Amsterdam: Benja- mins: 411-451. Poplack, Shana & Dion, Nathalie. (2004). The French future in grammar and speech. NWAVE 33. University of Michigan, Ann Arbor. Poplack, Shana, Dion, Nathalie, Jarmasz, Lidia-Gabriela, Leblanc, Carmen & Rosen, Nicole. (2002). Répertoire historique des grammaires du français. Corpus and documentation. Sociolinguistics Laboratory. University of Ottawa. 2000 pp. Poplack, Shana, Jarmasz, Lidia-Gabriela, Dion, Nathalie & Rosen, Nicole. (ms). The evolution of French prescriptive discourse: constructing the Répertoire historique des grammaires du français. Technical Report. Sociolinguistics Laboratory. University of Ottawa. 40 pp. Poplack, Shana & St-Amand, Anne (2002). A real-time window on 19th century ver- nacular French: the Récits du français québécois d'autrefois, CLA, University of Toronto. Rand, David & Sankoff, David (1990). GoldVarb. Version 2. A Variable Rule Appli- cation for the Macintosh, Centre de recherches mathématiques, Montréal: Uni- versité de Montréal. 14 MARTIN ELSIG & SHANA POPLACK

Roberts, Ian (1993). Verbs and Diachronic Syntax. A Comparative History of English and French, Dordrecht, Boston, London: Kluwer Academic Publishers. Söll, Ludwig (1982). L'interrogation directe dans un corpus de langage enfantin. In Hausmann, Franz Josef (ed.), Études de grammaire française descriptive, Hei- delberg: Groos. Terry, Robert (1970). Contemporary French Interrogative Structures. Montréal: Éditions Cosmos. Vaugelas, Claude Favre de (1880/1884). Remarques sur la Langue Françoise: com- prenant le texte de l'édition originale, des remarques inédites, une clef inédite de conrart, tous les commentaires du XVII siècle, des notes nouvelles, une introduction et une table analytique des matières, Paris: A. Chassang, 2 volumes.

Martin Elsig Collaborative Research Center: Multilingualism University of Hamburg D-22765 Hamburg, Germany [email protected]

Shana Poplack Department of Linguistics University of Ottawa Ottawa, Ontario K1N 6N5 [email protected]